Significant adverse effects on the course of embryo development have been reported in several mammalian species after exposure to high levels of "noise" during pregnancy in some, but not all studies conducted. An investigation was conducted to determine whether the characteristics of different noise exposure paradigms contributed significantly to this variability. Accordingly, three differing paradigms, thought to be representative of a wide range of noises, were devised to simulate (1) semi-continuous exposure to extremely high levels (126 dBA jet engine noise), (2) unanticipated high intensity starting sounds (microprocessor controlled pseudorandom onset, fast rise time alarm bells, sonalert type warning devices, or jet engine noise each at 112 dBA), (3) very high frequency sound to compensate for the upward frequency shift in rodents' hearing relative to humans (18-20 kHz modulated times at 110 dB from a commercial rodent repelling device). Pre- and post-implantation exposure to the first two paradigms resulted in decreased pregnancy maintenance and embryo lethality. Postimplantation exposure resulted in significantly increased fetolethality. No exposure related changes in plasma corticosterone levels were found.